the collaboration of OATs with ESO

SAIt 2012 c Memorie

VLT instrumentation control:

the collaboration of OATs with ESO

P. Santin

Istituto Nazionale di Astrofisica – Osservatorio Astronomico di Trieste, Via Tiepolo 11, I-34131 Trieste, Italy, e-mail: santin@oats.inaf.it

Abstract. The collaboration of the Astrophysics Technology Group (ATG) of the Astronomical Observatory of Trieste with ESO will be presented. It may be considered as an example of a good practice in the collaboration of a peripheral Institute with ESO.

Started in 1985, it spans almost all the time since Italy joined ESO in 1982. The Trieste Astrophysics Technology Group (ATG) participated in many technological enterprise of ESO, from Remote Observing to VLT instrumentation and to ALMA ACS, and is now looking for the challenging E-ELT.

Key words. Instrumentation: Control

1. Introduction

The collaboration of the Astrophysics Technology Group of the INAF - Astronomical Observatory of Trieste spans 27 years, almost the entire period of thirty years of participa- tion of Italy to ESO, that will be celebrated during this Workshop. I believe that this may be considered as an example of a good practice in the collaboration of a relatively small group of a peripheral Institute with the Technical Divisions of ESO, in the field of state-of-the-art technology applied to Astrophysics.

Let’s start with some dates : Italy joins ESO in 1982, and the Astrophysics Technology Group was formed in Trieste in 1975. We im- mediately recognized the need of strict con- tacts with the Technology Divisions of ESO, to stay connected with the state-of-the-art tech- nology in Astrophysics, and we started a pre- liminary collaboration ( still continuing today).

Send offprint requests to: P. Santin

The technological driver of this long-lasting and fruitful collaboration throughout the whole period has ever been the participation to the projects involving the periodic ESO innova- tions in the fields of Instrumentation Control Software, Instrumentation Control Electronics and Data Processing Software, giving always attention to the cross-section with the scientific interests present at OATs.

The different projects in which ATG has been involved in this 27 years will be pre- sented, in chronological order, with an intro- duction also to a couple of collaborations be- fore the VLT era.

2. Pre-VLT era - Image Display (1985 - 1988)

In the first years of the eighties ESO started to

develop MIDAS (Munich Image Display and

Analysis System) as a tool to handle the as-

tronomical data produced by the instrumenta-

tion at La Silla. The image display was imple-

Astronet for the definition of the IDI, Image Display Interfaces, a set of standard software interfaces to match the existing image display software systems to the new class of devices on the market. ATG was representing Astronet.

In 1987 the IDI were finalized and published (Terret et al. 1988) and ATG started a collab- oration with the ESO MIDAS group for the implementation of the IDI for MIDAS. The portable MIDAS system was born (pMIDAS), with image display capabilities available for a large number of devices, and it was possible from that time for ESO to distribute it to the pe- ripheral astronomical institutes that could cre- ate local DPCs.

3. Pre-VLT era - Remote Observing (1989 - 1992)

At that time the Remote Observing mode was already active at La Silla Observatory, operat- ing with the CAT. ESO decided to go for a re- furbishment and an upgrade of the system to be installed on the NTT. In the framework of a direct collaboration ATG is in charge of the implementation of the Client side (Garching client), and it is agreed that a feasibility study should also be started to verify the possibility to extend the Remote Observing capabilities to a different astronomical institute in Europe.

In a three years effort the Garching Client has been completely rewritten. Moreover the Second Level Remote Observing system was implemented, providing observing capabi lities to any European astronomical institute. A sec- ond level client has been implemented, to be installed in the remote institute, together with a remote observing manager, running at ESO

Fig. 1. Layout of the Second-level Remote Observing system from La Silla

HQ in Garching, whith the full control of the operations (Balestra et al. 1992). See Fig. 1 for the geographical deployment.

Those were not times of Internet, wide band, fast communications and a complex hardware system has been designed and set up, providing a communication over three chan- nels for data, video and voice, with a ded- icated link of 64KB between Garching and Trieste(see Fig. 2).

At the end, after three years of implementa-

tion and tests, three nights have been allocated

with NTT and 10 Italian astronomers from dif-

ferent Observatories observed with NTT and

SUSI for a total of more than 30 hours of ob-

serving time (Balestra et al. 1992).

Fig. 2. Harware architecure of the Second-level Remote Observing system from La Silla

NOTE

In 2011, 19 years later, in the framework of the EVALSO EU FP7 project (partners were ESO, GARR, UniTS, CLARA, REUNA, QMUL, AIRUB, INAF OATs) the scheme was repli- cated on the VLT (4). In 2011 the Remote Control was implemented via a pre-loaded OB with P2PP from a remote institute, where the raw data were collected and a it was possible to modify and re-start the OB. The goals of the 1992 implementation were more ambitious, providing the full control of the telescope/instrumentation to the remote user (replicating the observatory user interface) and the availability of the raw data.

4. VLT - UVES Control Software (1992 - 1999)

With the advent of the VLT, Very Large Telescope, UVES (UltraViolet-Visual-Echelle- Spectrograph) was one of the first two in- struments build under complete responsibility (and funding) by ESO, together with ISAAC.

There was therefore no European Consortium in charge of its realization, and ATG was called by ESO Software Department for a direct col- laboration. The activity was really challenging, for it was a first step in the VLT project and the VLTSW (the Software framework for the con- trol of the VLT and of the instrumentation) was still in a development phase and we had in this way the possibility to participate to its com-

Fig. 3. UVES during the integration in Garching

pletion and tests. Moreover it was the moment of the transition of the VLTSW from the C programming language to the Object Oriented software technology and this fact was really stimulating.

In the preliminary agreement a feasibility study was included for the Data Reduction, which was closed after two years.

In the framework of the UVES project ATG delivered more than 10 man-years over a 7 years project. Figure 3 shows the instrument during its preparation in Garching, Figure 4 the ESO-ATG UVES team at Paranal and Figure 5 a snapshot of the celebration for the first light (September 27, 1999).

5. A strategic agreement

During the UVES realization a discussion started between ESO and ATG for the prepa- ration of a formal agreement for future collab- orations. At the end a document was prepared which stated (among other)

The Software Department in the Technical

Division (TSW) at ESO and the Astrophysical

Technologies Group (ATG) of the Astronomical

Observatory of Trieste (OATs) acknowledge

the long-standing and successful collaboration

Fig. 4. The UVES team in Paranal

Fig. 5. The UVES first light - September 27,1999

between the two groups in the field of astro- nomical instrumentation control. The purpose of this collaboration lies in the fact that on one side ESO can benefit of a qualified and pro- fessional team, with in depth knowledge and experience of the VLT Instrumentation and its control system, while ATG, on the other hand, will take advantage keeping in contact with the ESO technology, collecting experience to be used in the national and international ground based and space borne telescopes and instru- mentation programs in which the OATs is in- volved, involving at the same time young staff

software by the different Consortia and on the other the maintenance at Paranal Observatory of the running instruments. The scope was the implementation of a basic OS (the Observation Software control process), standard for all instruments, to customize according to the instrument architecture. BOSS was imple- mented, the Base Observation Software Stub, a standard Observation Software framework (Pozna et al. 2008). It is still a running project, and support all the instrumentation software. With the implementation of this soft- ware framework the standard common soft- ware is gaining in complexity (see Fig. 6), pro- viding on the other hand a decrease in the com- plexity of the Instrumentation Software imple- mented by the instrument teams (see Fig. 7).

Our team, through the implementation of the INS Software for many instruments collabo- rated (and is still collaborating) with the ESO responsible of BOSS for the improvement of the architecture and of the performance of the software package and for the complex tests on- the-field (Pozna et al. 2010).

Fig. 6. Complexity of the BOSS package during its

development

Fig. 7. Relative complexity of BOSS and of the INS software

7. VLT - FLAMES/GIRAFFE Control Software (1999 - 2002)

In 1999 ESO asked our team to participate in the realization of FLAMES/GIRAFFE. This instrument, a fiber-fed multi-object medium resolution spectrograph, was built by an European consortium plus Australian partners.

ATG was not part of the consortium but it came out that it was a complex instrument to be con- trolled and it represented an optimum test-bed for the exploitation of the Unified OS (BOSS) and of the SuperOS concept. From the control point of view it was composed by three instru- ments:

– OzPoz - a fibre positioning robot with a double plate for parallel configuration and observation (Fig. 8)

– GIRAFFE - a fiber-fed medium resolution multi-object spectrograph (Fig. 9)

– UVES - a high resolution spectrograph, al- ready operative on the opposite Nasmyth platform, fed with 8 dedicated fibers The layout of the instruments on UT2 is shown in Fig. 10.

The parallelism was the key issue of the in- strument operations. GIRAFFE and UVES can observe independently while the alternate plate of the Fibre Positioner is being configured for for the next observation.

The full parallelism was achieved with an architecture implementing

– three fully independent instruments, with a dedicated Observing software (BOSS) supervisor (BOSS-SuperOS) coordinating the operations

Fig. 8. The scheme of the fibre positioner of FLAMES

Fig. 9. OzPoz, the fibre positioner, and GIRAFFE (open)

– a double-BOB user interface for the Observation and Fibre Setup

The ATG team played a double role within this project

– a direct collaboration with ESO, under a specific agreement for the implementa- tion of the overall Control Software of FLAMES/GIRAFFE

– member of the ITAL consortium

(Bologna, Cagliari, Palermo and Trieste

Observatories) in the framework of

the FLAMES international consortium

(Australis, Paris-Meudon, Geneve and

Fig. 10. FLAMES and UVES on UT2

Lausanne Obs., ITAL) for the upgrade of the UVES control software.

The internal software of the robot of the fi- bre positioner OzPoz was implemented by the Australian partners. The man-power delivered for the whole project by the ATG team has been of 5 man-years.

8. VLTSW - Linux (2002 - 2003)

Linux was knocking at the door. The VLTSW was tied to a hardware architecture created around Unix HP workstations and VME elec- tronics. The Linux revolution was bringing a free, open-source software running on a wider choice of hardware, cheaper and with higher performances. ESO made available to ATG a preliminary beta version of the VLTSW run- ning under Linux for the first tests. During the last phases of the commissioning of FLAMES the sw has been unofficially installed on an Armada laptop under Linux and run in paral- lel with the official control workstation. The result has been an improvement of a factor of three in compilation performances and of a factor up to 10 in panel handling perfor- mances. ESO and ATG started a project to ver- ify the feasibility of a mixed environment with an Instrument Workstation running Linux. The Test Camera on the VLT was selected as a test-

After 15 years of VLT Software ESO, for the new great enterprise is moving towards a new software infrastructure, ACS ALMA Control Software. ACS provides a common software infrastructure to all the ALMA developers and creates a distributed environment for its software tools and services. ACS is based on the most advanced software technologies (e.g.

OO, C++, CORBA) for the management of its distributed objects. Thanks to a specific agreement with ESO and to the involvement of ATG in the development of ACS, the Trieste Observatory becomes an ALMA partner (see Fig. 11).

Fig. 11. The ALMA partners around the world

The collaboration covers several aspects of

ACS. In particular at OATs two main software

packages have been implemented and tested :

the Sampling System, to provide sampling of

the properties of the system for the instrument

telemetry, and the Bulk Data System, the core

of the data transfer from the antennas, provid-

ing a transfer rate of up to 800 Mbps.

10. VLT - X-shooter Control Software (2004 - 2009)

X-shooter, a medium resolution multi- wavelength spectrograph, is a second genera- tion instrument fully Linux based. Three arms, optimized for the Blue, Visible and near IR bands, were to be controlled, synchronized (to obtain fully parallelism) and keep aligned.

Most of the complexity of the software resides in the Observation Templates, where the full parallelism was implemented and where a complex real-time procedure was implemented to check and maintain the alignment of the three arms during the exposures. Before each scientific exposure two calibration exposure are taken, with different configuration of the instrument. The resulting images are then correlated to verify any displacement of the spectra. If the result of the correlation is good the observation template can then proceed.

The Trieste Observatory was part of the international X-shooter consortium (Denmark, Italy, Nederland, France, ESO) and for the Control Software of the instrument ATG deliv- ered 7 man-years (out of a grand total of 70 man-years for the full instrument).

11. VLT - ESPRESSO (2009 - ...2015) : the present

ESPRESSO (Echelle SPectrograph for Rocky Exoplanets and Stable Spectroscopic Observations), the current instrument, is a second generation VLT instrument and marks two important steps for the developers:

– it is the first VLT instrument to use the Combined Incoherent Focus. This means that it is a Coude instrument, installed in the CCL (Combined Coude Laboratory) which can combine the beam of up to four VLT UTs . An important task will be there- fore the upgrade of the VLT INS Common Software to use more than one telescope (that is the standard behavior for all the in- stalled instruments up to now.

– ATG/OATs is responsible in the case of ESPRESSO of the Instrument Control Software and of the Instrument Control Electronics.A big revolution is ongoing at

Fig. 12. The layout of the Coude Combined Laboratory with ESPRESSO

Fig. 13. The four VLT Unit Telescopes combining the beams

Fig. 14. ESPRESSO inside its container

ESO in the field of Control Electronics, abandoning the old LCUs based on a VME/VxWorks architecture and adopting a new modern architecture based on dis- tributed PLCs with the OPC-UA concept.

12. E-ELT : the future

In the framework of the Concept Studies

(Phase A) for the E-ELT Instrumentation

Fig. 15. E-ELT - the European Extreme Large Telescope

A first result of this work, discussing ad- vantages and/or disadvantages in terms of

bers)

I wish to thank all the ESO collegues in Garching and Paranal for their friendship and helpful collabo- ration.

References

Terret, D. L., et al. 1988, A&A suppl. 76 Balestra, A., et al. 1992, ESO Messenger 69 Balestra, A., et al. 1992, Proc. of Observing at

a Distance, Emerson and Clowes (eds.) Pozna, E., et al. 2008, Proc. SPIE 7019 Pozna, E., et al. 2010, Proc. SPIE 7740 Di Marcantonio, P., et al. 2009, ESO Technical

Report E-TRE-OAT-511-1053